Response Functions

Supporters of standard teleosemantics argue that informational teleosemantics turns teleosemantics on its head, because functions are effects but a representation’s information relations concern its causes. In chapter 6, the author responds to this influential objection by explaining that, while functions must involve effects, this is not to the exclusion of triggering causes. According to the etiological theory, which is employed by most proponents of teleosemantics, functions are (roughly speaking) selected effects; however, they can also be selected dispositions or selected causal roles, and so can involve inputs as well as outputs. The author explains that there are response functions (functions to do something in response to something), that sensory-perceptual systems have them, and so can have information-processing functions, at least given a simple causal analysis of information. This clears the path for the causal-informational version of teleosemantics, which ties the contents of (nonconceptual) sensory-perceptual representations to their normal causes, as opposed to the so-called Normal conditions for their use.

The Methodological Argument for Informational Teleosemantics

Teleosemantic theories are diverse, but they all endorse the claim that semantic norms, to do with correct and incorrect representation, derive in part at least from functional norms, to do with normal or proper functioning. Informational teleosemantics adds that semantic norms also derive from natural-factiveinformation. In this chapter, The author starts with the premise defended in chapter 3–– in explaining how bodies and brains operate, biologists use a notion of normal-proper function. To this the author adds that the same notion of function is used in explaining cognitive (including perceptual) capacities, and then argue that, given an information-processing approach, the norms of proper functioning are thus wedded to the aboutness of natural-factive information, so that a basic form of normative aboutness is posited. This elucidates the explanatory role of positing nonconceptual representations, establishes the scientific credentials of informational teleosemantics, and gives us good reason to try to solve its alleged problems. In the last few sections, the author argues that the main naturalistic “alternatives” to teleosemantics also have apparently ineliminableteleosemantic commitments.

Functional Analysis and the Species Design

Chapter 3 is about functional explanation in biology, rather than directly about mental content, but in this chapter the author defends a controversial premise of the methodological argument for teleosemantics (given in chapter 4). The premise is that physiologists and neurophysiologists ascribe normal-proper functions in explaining how bodies and brains operate for significant scientific reasons. How an organic system operates in the here and now depends on the actual causal contributions its components make in the here and now, and yet biologists also describe the normal-proper functions of components when explaining how (and not just why) complex organic systems operate or function the way they do. Central to the biologists’ task is describing systems that are functioning normally or properly (e.g., normal human visual systems, or normal human immune systems). The author explains how this role of a malfunction-permitting notion of function (sometimes called a “normative” notion) is consistent with the etiological theory of functions, but the aim here is not to establish the truth of the etiological theory of functions (which is defended in other works).

Positing Nonconceptual Representations

In the second chapter, the author describes some research by cognitive scientists, who posit nonconceptual representations to explain certain perceptual capacities (and incapacities). This research and the way in which it is reported illustrate the type of theoretical work done by an error-permitting notion of nonconceptual representation, alongside a malfunction-permitting notion of function. One set of studies (led by McCloskey) that is described in some detail focuses on an unusual deficit in locating visual targets (in a young woman, AH), which were intended to contribute to understanding normal human vision. The author makes clear why the contents ascribed to the underlying representational states, where the errors first occur, are referential-intentional contents, not merely (natural-factive) informational contents, and why their ascriptions count as intensional, according to standard criteria. Toward the end of the chapter, the author reminds readers of a familiar conundrum: if a representation’s having content is not causally potent in a psychological process, why is it (still) a central tenet of mainstream cognitive science that such a process should be understood as representational?

Distal and Distant Red Squares

The final chapter is about the notoriously difficult problem of distal content (the sixth of the content determinacy challenges listed in chapter 7). In relation to causal theories of reference, the problem is to explain why a representation counts as representing its content, rather than some other item that carries information about it to the sensory-perceptual system that produces the representation. After discussing how the problem concerns nonconceptual representations, as opposed to conceptualized thought, the author offers a solution suitable for nonconceptual representations. She explains why several well-known proposals do not seem to be even designed to assign appropriate distal contents to nonconceptual representations (whatever their success might or might not be for concepts). Neander alsoalso discusses several closely related issues, such as how a hallucinated red square might be experienced as “in the world,” versus “just in the head,” and how distal contents and perceptual constancies are related. Before concluding, the author discusses the issue of how informational teleosemantics handles the content of complex contents produced by the so-called “binding” of perceived features to perceived objects.

Causally Driven Analogs

In chapter 8, the author brings causal, teleosemantic and resemblance theories of content together by extending CT (as presented in chapter 7) to explain how homomorphism (more specifically, analog relations, or second-order similarity relations) can play a content-constitutive role. While the idea that some systems have the function to model the world might explain why a natural representational system counts as a representational system, it is neutral between iconic mental representation and a more language-like version of Mentalese. The author then addresses traditional objections to resemblance theories of content, to show how they can now be met. She argues that sensory-perceptual systems can have functions to produce inner state changes that are caused by and the analogs of their contents, and that this casts light on the intuitive appeal of resemblance theories of mental representation, and that it permits some sensory-perceptual simples to represent novel and non-existent contents. Chapter 8 also addresses the fourth and fifth content-determinacy challenges: why does R have the content there’s C and not there’s Q, when (iv) C is a determinate of Q, or (v) C is a determinable for Q? In the last few sections, Berkeley’s problem of abstraction and two contemporary strategies for its solution are discussed, insofar as it concerns the representation of perceptible properties (e.g., shape and color).

Simple Minds

Plainly enough, the right theory of mental content should (when fed the facts) assign the right contents to mental representations, but which ones are right? There is disagreement among philosophers especially regarding the relatively simple systems often used (and abused) as “test” cases. Some disagreement is due to different meta-analytic aims or background assumptions, but some is due to lack of familiarity with known facts. So, in this chapter, the author sketches some neuroethological findings regarding the vision involved in toad prey-capture. What is argued is that the information-processing approach to explaining the toad’s abilities imposes constraints on which contents ought to be ascribed in this case. The normal system, and its normal causal sensitivities, must be able to support the contents ascribed, and perceptible features must be represented prior to the representation of imperceptible ones (e.g., in vision, visible features must be represented before invisible ones). In exploring the implications of this for philosophical theories of content, the author argues that, if the information-processing approach is along the right lines, the right contents are not those delivered by standard (a.k.a. “High Church”) teleosemantics, such as Millikan’s.

Thinking about Thought

Thoughts are about items in (or aspects of) the world, and mental representations refer to them: how can this representational power be explained? In this chapter, the author introduces the philosophical problem of intentionality, clarifies how key terms will be used in this book, explains (and to some extent motivates) the author’s starting assumptions, and sets out her main aims. Among other things, she here discusses the naturalization project, relations between intentionality and consciousness, the distinction between original and derived intentionality, and the (natural-factive) notion of informational content as distinct from the notion of representational content involved in semantic evaluations. The aim in this book is to give a “modest” theory of original intentionality, a theory of referential-intentional content with its scope initially limited to sensory-perceptual (nonconceptual) representations. The focus will be on the origination question (what is the basis of original intentionality?), not on the derivation question (how do more sophisticated forms of intentionality derive from less sophisticated?).

The Content-Determinacy Challenges

In this chapter, the author begins developing the details of the causal-informational version of teleosemantics (CT) for (nonconceptual) sensory-perceptual representation. She also lists six content-determinacy challenges, and explain how CT handles the first three (leaving the others to be treated in chapters 8 and 9). Content determinacy challenges ask why a representation (R), which has the content C, has the content C (and not, say, Q). The first three challenges ask why R has the content there’s C and not there’s Q (given that C is not Q) and, (i) C and Q were co-instantiated where and when selection for the R-producing and R-using systems took place, (ii) C and Q were both causally implicated in past selection for the R-producing and R-using systems, and (iii) C and Q have necessarily been co-instantiated, even though “C” and “Q” do not co-refer. In relation to (iii) the author discusses the case of seeing green rather than grue, and the case of seeing a shape as a diamond rather than as a regular square. In effect, the author defendsStampe’s early idea, that taking the teleosemantic turn is a good way to improve upon a causal theory of reference.